Abstract
This study investigates the self-rationalization framework constructed with a cooperative game, where a generator initially extracts the most informative segment from raw input, and a subsequent predictor utilizes the selected subset for its input. The generator and predictor are trained collaboratively to maximize prediction accuracy. In this paper, we first uncover a potential caveat: such a cooperative game could unintentionally introduce a sampling bias during rationale extraction. Specifically, the generator might inadvertently create an incorrect correlation between the selected rationale candidate and the label, even when they are semantically unrelated in the original dataset. Subsequently, we elucidate the origins of this bias using both detailed theoretical analysis and empirical evidence. Our findings suggest a direction for inspecting these correlations through attacks, based on which we further introduce an instruction to prevent the predictor from learning the correlations. Through experiments on six text classification datasets and two graph classification datasets using three network architectures (GRUs, BERT, and GCN), we show that our method not only significantly outperforms recent rationalization methods, but also achieves comparable or even better results than a representative LLM (llama3.1-8b-instruct).
Abstract (translated)
这项研究探讨了基于合作博弈构建的自我合理化框架，其中生成器最初从原始输入中提取最具信息量的部分，随后的预测器使用选定子集进行其输入处理。生成器和预测器被协同训练以最大化预测准确性。在本文中，我们首先揭示了一个潜在的问题：这种合作博弈可能无意间在合理化抽取过程中引入采样偏差。具体来说，生成器可能会错误地建立所选理由候选者与标签之间的关联，即使在原始数据集中它们在语义上是无关的。 接下来，我们通过详细的理论分析和实证证据解释了该偏见的起源。我们的研究结果表明可以通过攻击方法来检查这些相关性，并基于此进一步引入指导以防止预测器学习到这些相关性。通过在六个文本分类数据集以及两个图分类数据集中使用三种网络架构（GRUs、BERT 和 GCN）进行实验，我们证明了该方法不仅显著优于最近的合理化方法，而且还实现了与代表性的大型语言模型（llama3.1-8b-instruct）相当甚至更好的结果。
URL
https://arxiv.org/abs/2505.02118
